Segmentation of Corporate Clients' Non-Performing Loans in Commercial Bank

In this study segmentation of small business unreliable borrowers was conducted to improve credit risk management system of the commercial bank «Sberbank». Based on tests for statistically significant differences between reliable and unreliable borrowers and also on the basis of the significance of potential risk factors in the constructed regression model of the probability of falling into non-performing loans factors for further segmentation were identified. Segmentation of bank's unreliable borrowers was conducted by the method of ABC analysis and the method of hierarchical cluster analysis. The segmentation made it possible to form groups of unreliable borrowers which divided by the degree of unreliability and to define the profiles of unreliable borrowers separately for individual entrepreneurs and limited liability companies. Between the segments of unreliable borrowers statistically significant differences were found in the variables of the loan amount, term loans and non-performing loans, the rate of falling into non-performing loans. During the analysis of the results obtained, the main risk factors which influence on falling into non-performing loans was identified, in particular, large loan amounts, small loan outstanding balance and the taking of credit during the crisis period (2014-2016). Together with the head of the audit department of corporate business, optimal measures for dealing with unreliable borrowers of different segments were identified. The results of the research can be used by the "Audit of corporate business" service of the bank to form a risk-oriented sample of borrowers for conducting audits in «Lending and dealing with problem assets of legal entities» department.